High-resolution data are improving our ability to resolve temporal patterns and controls on river productivity, but we still know little about the emergent patterns of primary production at river-network scales. Here, we estimate daily and annual river-network gross primary production (GPP) by applying characteristic temporal patterns of GPP (i.e., regimes) representing distinct river functional types to simulated river networks. A defined envelope of possible productivity regimes emerges at the network-scale, but the amount and timing of network GPP can vary widely within this range depending on watershed size, productivity in larger rivers, and reach-scale variation in light within headwater streams. Larger rivers become more influential on network-scale GPP as watershed size increases, but small streams with relatively low productivity disproportionately influence network GPP due to their large collective surface area. Our initial predictions of network-scale productivity provide mechanistic understanding of the factors that shape aquatic ecosystem function at broad scales.